Operating System Of Smart Home Based On Voice Interaction

The development of IT technology has driven all industries to develop in the direction of networking and intelligence.People's lives have also become more convenient and comfortable.One of the manifestations is the application of the Smart Home systems in people's real lives.The Smart Home integrates various terminal devices through a central management system to facilitate the intelligentization of home devices.Among them,human-computer interaction and device integration control are problems that need to be solved in Smart Home.They are the basis for the construction of smart home systems and,to a great extent,determine the performance of the system and the actual experience of users.Therefore,the use of speech recognition technology,Internetional technology,natural language processing and other technologies to build a high-performance,stable and convenient Smart Home system has great practical value.This article focuses on building a stable and convenient Smart Home system in voice control.The research content of this paper includes four aspects: the overall design of the intelligent home speech control system,the design of the speech recognition,the design of the terminal device control and the server-side design.The overall design of the system includes hardware and software.The hardware is based on mobile phones,TV box and smart speakers as the control hub,respectively,using the mobile phone microphone,Bluetooth remote control and microphone array to obtain user's voice instructions,using the SDK to integrate and contral smart terminal devices such as smart TVs and build ZigBee family network to control lightsor other non-intelligent equipment..The software includes a speech recognition module and a terminal control module based on Android.The cloud server provides the training of the speech recognition model and the processing of the dictation and recognition text,and then returns the training model parameters and the matched device control instructions.The speech recognition module includes speech signal input,speech wakeup,dictation recognition,command word recognition,and natural language processing.For the far-field recognition scenario in the Smart home,we use the microphone array to improve the audio quality.For the wake-up problem,the hidden Markov model based on wake-up words and phonemes is used in wake up verification and reduce the false wake-up rate.The command word recognition engine is established by building a command word table to implement basic home devices control;the communication and recognition engine is combined with server-side natural language processing to resolve complex user instructions and provide better services.The terminal control module is combined with Android technology to develop in a modular manner,which reduces system development complexity and facilitates integration and version updates,and we build a ZigBee home network for integrated control of non-intelligent devices which extends the application scene of the Smart Homes.The server side adopts the Nettybased C/S communication framework.The framework is designed in layers,includes persistence layer,service layer,and network interface layer.Its advantage is that it can handle large-scale concurrent requests and can meet the requirements of the control system for network services.